writing-plans
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using a Python-based utility () for project initialization, status verification, and management of document indices.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external specifications and requirement documents to generate executable implementation plans and verification commands.
- Ingestion points: External inputs such as Spec Briefs, Design Specs, and approved requirements are ingested via the workflows described in
SKILL.md. - Boundary markers: The skill does not specify the use of delimiters or protective instructions (e.g., "ignore embedded instructions") when interpolating external requirements into the planning context.
- Capability inventory: The skill has permissions to write implementation plans to the filesystem (
docs/aegis/plans/) and execute shell commands through the project's Python helper script. - Sanitization: No explicit sanitization, validation, or escaping of input requirements is performed to prevent malicious instructions from being incorporated into the resulting implementation tasks.
Audit Metadata